Jan-Paul, togethere with David Goy, Marketing Manager, Gaastra Online Shop, will run a session on Personalisation 2.0: Real-time purchase prediction and algorithmic visitor targeting for online shops at eCommerce Expo 2015. Join them to learn more.
When: 01 October 2015
Where: Olympia London
Today, most consumers already purchase anything online from electronics, fashion to insurance and financial products. There is less to gain from new visitors. Marketing efforts need to be increasingly tailored to meet the demands of existing and returning visitors. Management consultancy Bain & Company reports that a 5% reduction in the visitor defection rate, can increase profits by 5 – 95% and that it costs 6–7 times more to acquire a new customer than retain an existing one.
At the same time, behavioural targeting has become an essential asset for all ecommerce marketers. But only slowly do they start to apprehend that investing in behavioural targeting yields higher returns the later the retailers starts to engage with the visitor. The reason is simple: The later the engagement with a visitor, the more information is available about the characteristics, context and needs of the specific visitor.
Because later engagement with existing visitors holds much more potential for increasing success, optimising on-site experiences is an essential driver for any online business. Until now, digital marketers could generate quick wins with A/B and multivariate testing.
Online retailers test and select winning design options for static on-site experience. A multitude of simple and powerful software suites allow digital marketers to seamlessly optimise static appearances of their ecommerce outlets today.
However, A/B testing neglects a very important aspect in digital marketing especially in today’s world: Visitors are not alike and we have plenty of information at hand to understand their different contextual situations, characteristics and needs.
A/B testing does not acknowledge individual differences between visitors. It aims to optimise digital retail experience towards a lowest common denominator among visitors heterogenous needs and expectations.
Imagine that you have three different visitor personas on your site: Type A = 50% of visitors, Type B = 30% of visitors and Type C = 20% of visitors. Now, optimising your site based on A/B test results will lead to your site be predominantly rendered to fit the demands of type A visitors, and neglect the individual needs of Type B and Type C visitors.
The key to overcoming the limitations of A/B Testing is personalisation. Digital marketers need to start paying attention to the heterogeneity of visitor context and needs.
Personalisation branches out into many applications. Best known and most established are recommendation engines to recommend products that fit previous interest. On-site targeting technologies focus more on behavioural intentions during the visit such as intentions to exit the page or potential willingness to checkout baskets later that were filled but abandoned during the session. These technologies rely on simple heuristics for deciding which visitor to target. Typical examples for these heuristics are moving the mouse to close the tab, clicking on the tab close button, and abandoning the site during the checkout process.
While they have been tremendously successful in triggering one-time actions such as generating new leads, newsletter subscriptions or targeting first time visitors, these intent-based targeting technologies suffer tremendous drawbacks when engaging with recurring and existing visitors.
Simple intent triggers that offer coupons, rebates or special prices to new or returning visitors are quickly recognised and easily habitualised by users. In an amusing Barclays UK TV-Ad an older lady explains how online shoppers may save by staying intentionally inactive to receive a voucher. The retailer nudges his most valuable visitors to mimic the simple behavioural patterns to receive unnecessary rebates and coupons.
The key to effective personalisation is to focus on targets that are aligned with overall customer satisfaction. Long-term targets such as increasing customer lifetime value are more helpful than short-term session-based goals (“Create the conversion in this session no matter what!”) when implementing personalisation solutions.
Taking into consideration visitors’ cross-device visiting behavior, purchasing history, and previous site engagement allows retailers to only target visitors when the targeting will result in an increased customer lifetime value.
Digital marketers need to apply more complex targeting mechanics than session-based heuristics such as exit-intent targeting. Sophisticated prediction algorithms like random-forest models and sequential pattern mining need to be combined with intelligent targeting based on multi-armed bandit systems and intelligent message design to find the right message for the right moment. Ultimately, our own experience at Akanoo from large ecommerce clients who used exit-intent technologies before shows that more sophisticated algorithmic targeting yields more sustainable additional revenue and a large crowd of additional happy customers.
Wednesday, December 07, 2016